In measuring reliability and availability data, long periods of time are needed in order to get better and more accurate measurements. Often the amount of time needed in order to gather appropriate amounts of data for certain types of failures exceeds the amount of time that can realistically be gathered from a single computer. When possible this makes it desirable to be able to combine runtime information from multiple computers and be able to treat these groups of computers as a single system, thus ensuring that enough time has elapsed to more accurately measure reliability. To do this correctly, it is important to understand when it is appropriate to aggregate sets of computers into a single system and when grouping data from sets of computers not appropriate.